摘要
为防止受电弓滑板出现裂纹而造成重大弓网事故,针对电气化铁路受电弓滑板裂纹自动检测问题进行研究。算法以二代曲波变换为基础,通过研究曲波分解的方向性及各向异性特点,提出一种曲波系数定向映射(CCDP)算法。CCDP变换将分布在多个曲波分解角度矩阵中的高维信息变换到一个单一的二维矩阵中,从而将受电弓滑板上各种图形元素及其相互关系明确地显示出来。通过分析受电弓滑板CCDP变换矩阵,可区分并逐步排除滑板边界提取缺陷、滑板接缝以及滑板划痕的干扰,最终可判断滑板裂纹信息。实验证明,基于CCDP的受电弓滑板裂纹提取方法能够较好获得裂纹信息,可为完善受电弓视频在线监控系统提供一种方案。
The pantograph slide cracks detection technology is studied to avoid accidents aroused by pantograph slide cracks.On the basis of the 2nd generation curvelet transform,the new curvelet coefficients directional projection（CCDP） algorithm is proposed after analyzing directivity and anisotropy of the curvelet.The high dimension information distributed in different curvelet transform angle matrixes is projected to the single 2-dimension matrix by using CCDP transform,so that the image elements and the relationship between them are displayed explicitly.Through analyzing the CCDP transform matrix of the pantograph slide,the images of the slide edge partition defects,slide scratches and slide joints in the pantograph slide can be distinguished and excluded gradually so that judgement of slide cracks can be completed finally.The experiment shows that the detection method of the pantograph slide cracks based on CCDP transform is able to acquire better slide crack information and to provide a new solution for pantograph online monitoring.
出处
《铁道学报》
EI
CAS
CSCD
北大核心
2011年第11期63-69,共7页
Journal of the China Railway Society
基金
国家自然科学基金项目(51007074)
教育部新世纪优秀人才支持计划(NCET-08-0825)
教育部霍英东青年教师基金(101060)
四川省青年科技基金(07ZQ026-012)
