计算机工程与应用2019，Vol.55Issue(17)：9-19,11.DOI:10.3778/j.issn.1002-8331.1905-0255
图像直线段检测算法综述与测评
Overview and Evaluation of Image Straight Line Segment Detection Algorithms
摘要
Abstract
The existing line segment detectors are systematically reviewed and evaluated. In the overview part, the existing detectors are divided into two categories:the global Hough-transform-based method and the local perceptual grouping method. The technical basis and implementation mechanisms of these two kinds of algorithms are investigated, and their advantages and disadvantages are summarized. In the evaluation part, Cho et al. established the first objective evaluation system for line segment detectors in a recent work. Unfortunately, this evaluation system has serious bugs in both code implementation and practical application. In consequence, the results obtained in the previous work of Cho et al. have not given an accurate enough picture of the detection performance of the current algorithms. The bugs in the evaluation system are all corrected, and 8 state-of-the-art line segment detectors are evaluated and compared. Finally, the prospects of future trends in this field are discussed in order for facilitating the relevant research.关键词
直线段检测/Hough变换/感知组合/客观测评/主观测评/精度/召回率/F-值Key words
straight line segment detection/Hough transform/perceptual grouping/objective evaluation/subjective eval-uation/precision/recall/F-score分类
信息技术与安全科学引用本文复制引用
郑行家,钟宝江..图像直线段检测算法综述与测评[J].计算机工程与应用,2019,55(17):9-19,11.基金项目
国家自然科学基金(No.61572341) （No.61572341）
江苏高校优势学科建设工程资助项目. （）
