机车全动态包络线测量中的目标点自动识别算法研究  

Research on Automatic Recognition Algorithm of Target Points in Locomotive Dynamic Envelope Measurement

作　　者：邹志[1] 马骊群[1] 甘晓川[1] 樊晶晶 ZOU Zhi;MA Liqun;GAN Xiaochuan;FAN Jingjing(Changcheng Institute of Metrology&Measurement,Beijing 100095,China)

机构地区：[1]航空工业北京长城计量测试技术研究所,北京100095

出　　处：《计测技术》2018年第5期43-47,共5页Metrology & Measurement Technology

摘　　要：针对机车动态限界图像中目标点自动识别的问题,提出一种基于双分辨力分析的目标点识别算法,解决了高分辨力图像对于多目标的搜索效率问题,并提出了一种自描述向量用于改进传统的RANSAC算法,以减少背景噪声的干扰,完成了不同限界图像目标点的单应矩阵的求解问题。最终实现高效且准确的目标点自动识别与匹配。For the problem that recognized the target points in the dynamic gauge image,a method of target auto recognization based on dual resolution analyzation is proposed,it accelerate the efficiency of the multi-target searching problem in high-resolution image.The author put forward the self-describedvector,to improve the traditional RANSAC algorithm.The self-described vector can remove the interference from the background noise,and solve the homography of different gauge images.Automatic recognition and matching of target points can be achieved.

关 键 词：动态限界 自动识别 双分辨率分析 自描述向量 RANSAC 

分 类 号：TB9[一般工业技术—计量学] TP273.3[机械工程—测试计量技术及仪器]