改进FDSST算法在铁路侵限异物检测场景中的应用  

Application of Improved FDSST Algorithm in Railway Intrusion Foreign Object Detection

作　　者：陶正轩 王小鹏[1] Tao Zheng-xuan;Wang Xiao-peng(School of Electronic and Information Engineering,Lanzhou Jiaotong University,Lanzhou 730070,Gansu Province,China)

机构地区：[1]兰州交通大学电子与信息工程学院,甘肃兰州730070

出　　处：《科学与信息化》2025年第2期169-171,共3页Technology and Information

摘　　要：针对铁路异物跟踪检测算法在实际应用中存在如异物目标存在遮挡、快速移动等因素影响,导致目标跟踪失败及漏检等问题,提出了一种改进FDSST铁路周界异物跟踪检测方法,利用子空间重构跟踪器和改进动态L1-PCA方法构建并维护跟踪模型的子空间,克服了遮挡和快速移动的影响;引入模型动态更新策略,解决跟踪失败问题并提升运算速度。实验证明,该方法准确率优于传统方法,有效提高了铁路异物检测性能。In practical applications,railway foreign object tracking and detection algorithms face challenges such as occlusion and rapid movement of the target,leading to tracking failure and missed detections.To address these issues,an improved FDSST(Fast Discriminant Scale Space Tracking)algorithm for railway perimeter foreign object tracking and detection is proposed.This method utilizes a subspace reconstruction tracker and an improved dynamic L1-PCA(Principal Component Analysis)approach to construct and maintain the subspace of the tracking model,overcoming the effects of occlusion and rapid movement.Additionally,a model dynamic update strategy is introduced to resolve tracking failures and enhance computational speed.Experiments demonstrate that this method outperforms traditional approaches in accuracy,effectively improving the performance of railway foreign object detection.

关 键 词：铁路侵限异物 FDSST算法 动态L1-PCA 模型动态更新 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]