针对现有铁路异物检测算法无法实时检测的问题，提出了一种基于 FPGA 的铁路异检测算法的硬件实现。该算法实现了图像的采集、运动目标检测、单次扫描连通域标记和异物特征提取，设计二维数组的存储结构用于记录标号、等价关系和特征参数，在单次扫描期间完成异物多个特征参数的存储，扫描结束后完成参数的整理和提取。最后，在搭建的铁路异物检测硬件平台上实现了该算法，并在此基础上进行了验证试验和现场试验。试验结果表明，该算法能够有效提取检测异物特征，参数提取正确，速度满足实时检测的要求，可在铁路现场完成异物的实时检测，多参数的提取可用于异物的分类和跟踪。%Aiming at the problem of existing railway clearance detection algorithms of being unable to realize re-al-time detection,a hardware implementation of railway clearance detection algorithm based on FPGA was pro-posed in this paper.This algorithm realized the image acquisition,moving object detection,single pass con-nected component labeling and clearance feature extraction.A storage structure of two-dimensional array was designed to record labels,equivalence relations and feature parameters of clearance.During the single pass,the storage of multiple feature parameters of clearance was completed.After the pass,the collation and extraction of parameters were completed.Finally,the algorithm was implemented on the hardware platform of railway clearance detection and verification and field experiments based on the algorithm were conducted.According to the experiment results,the algorithm can deliver effective extraction and detection of the characteristics of clearance and accurate parameters extraction with the processing speed meeting real-time detection require-ments.It can be applied to the real-time detection of railway clearance and multi-parameter extraction can be used for classification and tracking of clearance.
展开▼
