摘要
为了解决传统直线提取算法中由于梯度信息孤立而造成检测效果不理想的问题,本文提出结合边缘块的连通域信息进行边缘连接的直线提取方法.首先提取图像边缘图,利用边缘点的连通域,建立边缘点标注模型,进行边缘块搜索.然后用邻接矩阵统计边缘块之间的位置信息,采用深度优先搜素算法,得到初始路径.最后利用直线的几何特征作为路径筛选条件,完成图像直线特征检测.通过理论分析与实验研究表明,该算法可以有效地完成各类场景中的直线特征提取.
Line Detection is a basic problem in Computer Vision.When the gradient information is isolated,the detection effect is not ideal in the traditional line extraction algorithm.To address the issue of gradient information,we propose a novel line extraction algorithm base on edge block which capture the connected domain information of edge blocks to link edge map.Firstly,the edge graph of the image is extracted.And then labeling model is constructed by analyzing the connected regions properties of edge points,which is used to search edge block.In this way,point features are aggregated into block features.Next,calculating position information between edge blocks by adjacency matrix,and using the depth-first search algorithm to obtain the initial path.Finally,the geometric features of line are considered as the path screening condition to complete line detection.The experimental result showns that our algorithm can effectively extract line segments in various scenes.
作者
申超胜
张敬峰
林靖宇
SHEN Chao-sheng;ZHANG Jing-feng;LIN Jing-yu(College of Electrical Engineering,Guangxi University,Nanning 530004,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2021年第4期773-778,共6页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61561005)资助。
关键词
直线检测
边缘块
邻接矩阵
深度优先搜素
路径
line segment detection
edge block
adjacency matrix
depth-first-search
path
