基于线特征的影像与矢量的匹配

Line feature-based matching between imagery and vector map

摘要:

在影像与矢量匹配中，矢量数据是匹配的依据。其中，有些地物矢量是以边缘线形式出现的；另一些则表现为地物中心线。为了保证影像线特征提取的结果与矢量数据的一致性，本文提出了基于对称小波断面检测法和基于OTSU分割的形态算子检测算法。对称小波断面检测法主要用于提取影像中地物的边缘线；基于OTSU分割的形态算子检测法主要用于提取影像中地物的中心线，为影像与矢量匹配提供基础。同时，针对GIS矢量数据极适合用线矩来描述的特点，研究了3参数匹配的方法，这一方法中三个参数具有旋转不变性，能够得到可靠的匹配结果。

Abstract:

By matching the image to vector map, the vector data is the reference. Thereinto, some vector of ground object are edge line and others are center line. It is necessary to keep the consistency between the result of image line-feature extraction and vector data. Here two kinds of detection algorithm are put forward. The first one is based on symmetric wavelet and applied to extract the edge line in the image. The second one aims at detecting the center line of the ground object from image through OTSU segment and morphologic processing. which are the basis of matching between imagery and vector map. Since the line moment is suitable for expressing vector data, we have studied one matching method with three rotation-invariant moment parameters. This method brings more effective and better matching results.