详细信息
文献类型:期刊文献
中文题名:基于SIFT特征的铁路扣件状态检测算法
英文题名:Railway fastener state detection algorithm based on SIFT feature
作者:赵珊珊[1];何宁[2];曹珊[1]
第一作者:赵珊珊
机构:[1]北京联合大学北京市信息服务工程重点实验室;[2]北京联合大学智慧城市学院
第一机构:北京联合大学北京市信息服务工程重点实验室
年份:2018
卷号:37
期号:11
起止页码:148-150
中文期刊名:传感器与微系统
外文期刊名:Transducer and Microsystem Technologies
收录:CSTPCD;;CSCD:【CSCD_E2017_2018】;
语种:中文
中文关键词:铁路扣件;SIFT特征;Fisher向量;线性分类器
外文关键词:railway fastener;SIFT feature;Fisher vector;linear classifier
摘要:铁路扣件是铁路轨道线路的关键部位,其状态直接影响行车的安全。针对高速扫描相机采集的铁路数字图像,提出了铁路扣件状态检测算法。利用SIFT特征提取算法提取每幅扣件图像的局部特征,将每幅扣件图像提取到的SIFT特征归一化为相同长度的Fisher向量,使用LIBLINEAR分类器对归一化的Fisher向量进行分类,从而实现扣件图像的状态检测。实验结果表明:提出的铁路扣件状态检测算法具有较高的识别率及鲁棒性。
Railway fasteners are key components of railway tracks,and their condition directly determines the safety of the trains passing over them.A railway fastener state detection algorithm for railway digital image acquired by high-speed scan camera is presented.Using the SIFT feature extraction algorithm to extract local features of each fastener image,the SIFT features extracted by each fastener image is normalized to Fisher vector with the same longer.Use the LIBLINEAR classifier to classify Fisher vector,to realize state detection of railway fastener image.Experimental results show that the proposed algorithm has a higher recognition rate and robustness.
参考文献:
正在载入数据...