一种基于边界斜率拟合的角点检测方法  

A slope fitting corner detection algorithm based on edge chain-code

作　　者：黄天云[1] 姚远 HUANG Tian-yun;YAO Yuan(School of Mathematics,Southwest Minzu University,Chengdu 610225,China)

机构地区：[1]西南民族大学数学学院,四川远成都610225

出　　处：《西南民族大学学报（自然科学版）》2024年第4期428-435,共8页Journal of Southwest Minzu University(Natural Science Edition)

摘　　要：图像中的角点为描述物体特征提供关键信息,是复杂应用(如图像分类、目标检测和跟踪、定位和测量)的预处理步骤,角点检测的质量将直接影响后续图像处理的有效性.在工业环境中,角点检测算法需要在各类噪声和干扰因素下,对大规模数据集进行高效处理,以实现实时和准确的角点检测.因此,研究和设计快速高效、高准确性的角点检测算法具有重要意义.针对传统算法需要进行曲率计算或曲线拟合的局限性,提出了一种基于Freeman边界链码的快速、轻量级角点检测算法,通过对Freeman链码在角点之前和之后的连续多个点进行线段的斜率拟合和夹角的阈值判定,进而识别出角点.从准确性、鲁棒性和计算速度等方面,在NRS工业图像集上与主流角点检测算法进行了对比实验.结果表明,所提出的算法具有较少漏检和误检的角点数量,并实现了更快的检测速度,在工业应用中更具有优势.The corner points in an image provide key information for describing object features.Corner detection is the pre-pro-cessing step for complex image applications,and its quality directly affects the effectiveness of subsequent steps.In industrial environment,corner detection often requires real-time,accurate and efficient processing on large-scale image dataset while facing various types of noise and interference,so design high efficient corner detection algorithm is significantly important in prac-tice.Focusing on the limitations of traditional corner detection algorithms that require curvature calculation or curve fitting,a fast and light-weight method based on Freeman chain coded edge points was proposed,and whether the edge point was corner could be determined by slope fitting on multiple consecutive points before and after the current edge point and threshold on the includ-ed angle.The simulations were conducted in NRS industrial image dataset,its performances were evaluated in terms of accuracy,robustness,and speed,as compared with traditional algorithms such as HARRIS,SUSAN and SIFT,etc.The results showed that the proposed algorithm had fewer false positives and negatives with faster detection speed,exhibiting more superiority in indus-trial applications.

关 键 词：角点检测 特征提取 FREEMAN链码 工业应用 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]