一种基于边缘直线检测的矩形提取方法
2017-12-11卜飞宇
摘要:在图像识别中,矩形作为最常见的人工图形之一,对它的检测有着广泛应用。提出了一种基于边缘直线检测的矩形提取方法。该方法首先对图像进行边缘检测,得到边缘图像;然后对边缘图像进行直线检测,得到图像中所有的直线段;最后对满足矩形特征的直线段进行组合,提取图像中可能的矩形。实验结果表明,该方法具有较好的抗干扰能力,且运算速度快,是一种有效的矩形提取方法。
关键词:矩形提取;图像识别;边缘检测;直线检测
中图分类号:TP391 文献标识码:A 文章编号:1009-3044(2017)31-0182-03
An Approach to Rectangle Extraction Based on Edge Line Detection
BU Fei-yu
(Institute of Information and Electronic Engineering, Hunan City University, Yiyang 413000, China)
Abstract: In the image recognition, rectangle as one of the most common man-made shapes, detection of it has a wide application. An approach Based on edge line detection to extract rectangles in images is presented in this paper. The method firstly applies edge detection on the input image and obtains the edge map; secondly, it performs line detection to the edge map and obtains all straight line segments; thirdly, the line segments that satisfy the rectangular feature are combined to extract possible rectangles in the image. Experimental results show that the method has better anti-jamming ability, and it is fast and effective.
Keywords: rectangle extraction; image recognition; edge detection; line detection
1 概述
目標检测是图像识别中一个非常关键的任务。矩形是最常见的人工图形之一,对它的检测在图像识别系统中不可缺少。如车牌识别、广告牌检测等计算机视觉应用;从卫星遥感图像或飞机航拍图像中提取建筑物、车辆目标等。另外,矩形标识可用作无人机、智能机器人的定位地标,因此矩形检测可以用于机器人导航。
矩形检测的深入系统研究是最近若干年才开始的。一种矩形检测方法[1]是用Canny边缘检测算法得到线段,然后从这些线段中检测出T型和L型连接,将它们分解成更小的片段,再将小的片段进行联合线性化组成更长的线段。Tao等人[2]则将轮廓跟踪和直线分割技术应用到边缘图像中,由此获得图像中的直线段,然后通过分析和合并直线段来构造矩形。Jung和Schram提出了一种窗口化的霍夫变换[3],先运用霍夫变换获得候选直线,然后通过检测候选直线的组合能否满足矩形性质来检测矩形,但该方法中的一些参数需要调试。……
