基于改进图像结构算法的钣金件特征点识别  

Sheet Metal Feature Points Recognition Based on Improved Image Structure Algorithm

作　　者：徐杰 邹媛媛[1,2] 郑家俊 XU Jie;ZOU Yuanyuan;ZHENG Jiajun(School of Mechanical Engineering,Shenyang Jianzhu University,Shenyang 110168,China;High-Grade Stone CNC Processing Equipment and Technology National and Local Joint Engineering Laboratory,Shenyang 110168,China)

机构地区：[1]沈阳建筑大学机械工程学院,沈阳110168 [2]高档石材数控加工装备与技术国家地方联合工程实验室,沈阳110168

出　　处：《组合机床与自动化加工技术》2023年第7期176-179,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：辽宁省自然科学基金项目(20180551124)。

摘　　要：针对平面钣金件轮廓特征点识别中,特征点受边缘噪声影响,易出现伪特征点、切点难识别及特征点识别定位精度低等问题,提出了一种基于改进图像结构算法的平面钣金件轮廓特征点识别方法。首先,对特征点进行粗提取,通过判断轮廓点两相邻轮廓段的类型,并计算轮廓点间斜率差的绝对值,将斜率差的绝对值和阈值比较,分别获取切点和角点的候选特征点集;其次,进行特征点精确识别,将候选特征点进行分组融合,并引入直线度指标对线段轮廓上的伪特征点进行滤除,最终完成特征点的准确识别。实验结果表明改进的识别算法可以准确地识别平面钣金件轮廓特征点,能够滤除掉所有伪特征点,识别精度更高。Recognition of contour feature points for planar sheet metal parts are affected by edge continuity and noise,and are prone to problems such as fake feature points and low locating accuracy.In this paper an improved contour feature points recognition method of planar sheet metal parts based on the image structure algorithm is proposed.Firstly,feature points are coarsely extracted.By identifying the type of two neighborhood primitive of a contour point,and calculating the absolute value of the slope difference between the contour points,the candidate feature points set of tangent points and corner points will be obtained respectively after comparing the absolute value of the slope difference with the threshold.Then,accurate feature points are identified.The candidate feature points are merged and the straightness index is introduced to filter out the fake-feature points on the line primitives Finally complete the accurate identification of the feature points.The experimental results indicate that the improved recognition algorithm can filter out all fake feature points and recognize the feature points of the planar sheet metal contour with high accuracy.

关 键 词：特征点识别 平面钣金件 图像结构 直线度 

分 类 号：TH16[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]