融合角点信息的直线增强

直线提取是图像理解和识别的基础,相位编组法是常用且较为快速的直线提取方法之一。传统的相位编组方法存在着诸多问题:如一条直线容易断裂成若干段,直线交点及其附近直线上的点检测不到,无法得到物体的封闭轮廓,造成对物体描述与识别的困难。本文对相位编组方法存在的问题对其进行改进,提出一种利用角点信息对相位编组方法提取的直线进行修正,同时利用相交直线信息对角点位置进行修正的方法。

Straight line enhancement by using corner information

Straight lines play a very important role in image understanding and recognition. There are many approaches to extract straight lines such as Hough Transformation method, but they all have their particular drawbacks. Although the phase- grouping method is an effective algorithm, there are some limitations. First, a line in an image could be detected in several detached segments. Second, usually it is difficult to extract lines near corners or junctions. An algorithm of enhancing straight line extraction by using corner information is proposed in this paper. The positions of comers can also be adjusted accordingly. Experiments on both synthetic and real images show that the proposed method can produce more complete straight lines.