基于支持向量机与尺度不变特征转换算法相结合的线状矢量数据匹配方法

Matching Method of Linear Vector Data Based on Support Vector Machine and Scale-invariant Feature Transform Algorithm

由于不同来源或不同时间,导致了同一地物在存储方式或属性等方面存在着很多差异,这给后期的数据处理和使用带来了诸多不便,因此矢量空间数据匹配已经成为关键性问题。为了便于矢量数据匹配,提出了一种基于栅格化的线状矢量数据匹配方法,其主要思想是对线状矢量数据栅格化,利用支持向量机(support vector machine,SVM)算法提取出所要研究的数据,然后利用Harris算子提取特征点,用尺度不变特征转换(scale-invariant feature transform,SIFT)算法计算特征向量并对其进行匹配,最后把匹配结果转换为矢量数据。结果表明该算法不受平移、旋转、缩放、明亮度变化等的影响,弥补了矢量匹配过程中因数据旋转等问题而无法匹配的不足,将矢量数据栅格化处理,使其数据结构更简单,操作容易,更易于算法的实现。可见该方法便于线状矢量数据的匹配,为道路匹配提供了一种新的方法。

Due to different sources or different times,there are many differences in storage methods or attributes,which brings a lot of inconvenience to the later data processing and use. Therefore,vector space data matching has become a key problem. In order to make vector data matching more convenient,a matching method of linear vector data based on rasterization was proposed. Its main idea was to rasterize linear vector data,used support vector machine algorithm to extract the data,and then used the Harris operator to extract feature points,used scaleinvariant feature transform algorithm to compute the feature vectors and matched them,and finally converted matching results into vector data. The algorithm is not affected by translation,rotation,scaling,brightness change,etc.It compensates the lack of matching due to data rotation and other problems in the vector matching process. The vector data is rasterized to make its data structure simpler and easier to operate and easier to implement algorithm.It can be seen that this method is convenient for the matching of linear vector data and provides a new method for road matching.